What a litigation-ready summary looks like.

A medical record summary condenses hundreds or thousands of pages of treatment records into a short, cited brief a legal or claims team can act on. This page breaks down sample summaries section by section: what a strong summary includes, 4 common formats, and a template you can adapt.

What a strong summary includes.

Formats vary by firm and case type, but summaries that hold up under deposition share the same skeleton. Every section below exists to answer 1 reviewer question: can I trust this statement without re-reading the record?

Case header & referral questionWho the claimant is, what the matter is, and the specific question the summary answers. If the question isn't written down, the summary can't be judged against it.

Source inventoryEvery packet and page range reviewed, plus what was excluded and why. Duplicates and wrong-patient material get logged, not silently dropped.

Treatment chronologyVisit-by-visit entries in date order, each tied to its source page. This is the backbone the rest of the summary hangs on.

Key findingsThe material facts the case turns on, stated plainly, each with a page citation a reviewer can jump to and verify.

Work & functional statusRestrictions, capacity, and return-to-work history over time: the section adjusters and IME reviewers read first.

Flags for the reviewerGaps in treatment, inconsistencies between providers, unreadable pages, and anything quarantined. A summary that hides its uncertainty is a liability.

4 sample formats, and when each fits.

There is no single correct format. There is a correct format for the task in front of you. These 4 cover most litigation and claims work; many teams pair a narrative summary with a full chronology on larger files.

Format	Best for	Watch out for
Narrative summary	Demand support, IME referrals, and any reader who needs the story of the treatment in prose.	Drift: without per-statement citations, narrative invites paraphrase that the record doesn't support.
Tabular summary	Claims review and multi-provider files where the reader scans for dates, providers, and findings.	Context loss: a grid can't explain causation or why a gap matters. Pair with a short narrative lead.
Chronology-led	Complex treatment histories where sequence is the argument: what was known, when, by whom.	Length: a full chronology is a working document, not a brief. Experts and executives still need the summary layer.
Hybrid	Most real files: a 1-page narrative up top, a tabular or chronology body, flags at the end.	Duplication: keep each fact in one place with 1 citation, or versions will disagree.

A sample summary, section by section.

Read a sample the way a reviewer would: top to bottom, checking that each section does its 1 job. The table below walks the standard structure and what a reader should find in each part of a well-built summary.

Summary section	What the sample should show
Case header & referral question	Claimant, matter type, and the question the summary answers, stated in 2 lines at the top. Every later section is read against this question.
Source inventory	Each packet with page ranges, plus exclusions with reasons, so the reader knows exactly what was reviewed and what wasn't.
Treatment chronology	Dated, visit-by-visit entries, each cited to the page it came from. Sequence gaps should be visible, not smoothed over.
Key findings	The handful of facts the referral question turns on, each with a jump-to citation. If a finding has no page number, it doesn't belong here.
Work & functional status	Restrictions and capacity over time, with the source of each entry. Watch for status changes with no documented event behind them.
Flags for the reviewer	Everything uncertain, missing, or excluded, surfaced instead of buried: treatment gaps, provider disagreements, unreadable or quarantined pages.

Why quarantine logging matters: a single wrong-patient page contaminates every conclusion drawn after it. Catching that page before review, and logging it rather than deleting it, is the difference between a defensible summary and one that falls apart when opposing counsel finds the page.

A summary template you can adapt.

Field names differ across firms; the structure below covers what a litigation-ready summary needs. The point is not the exact fields. The point is that the outline exists before the deadline does, so nothing gets skipped under pressure.

Case & patient identifiersMatter number, claimant, date of loss, jurisdiction, and the reviewing team.
Referral questionThe specific question this summary answers, in 1 or 2 sentences.
Records reviewedPacket inventory with page ranges, sources, and date spans.
Exclusions & quarantine logDuplicates removed, unreadable pages, and any wrong-patient material, with reasons.
Treatment chronologyDated entries, provider, event, and a source-page citation per line.
Key findingsMaterial facts with citations, ordered by relevance to the referral question.
Work & functional statusRestrictions, capacity evaluations, and return-to-work history over time.
Gaps & inconsistenciesTreatment gaps, provider disagreements, and unexplained changes in course.
Reviewer sign-offWho verified the draft, what they checked, and when: the human gate, in writing.

AI drafts. Humans decide.

Software is good at the parts of this work that punish humans: sorting packets, removing duplicates, catching wrong-patient pages, and drafting against a template with a citation on every statement. The judgment stays with your reviewer. Medrecords AI does not retrieve records from providers. It reviews and summarizes the records you already have. It never produces signed opinions, case-merit scores, or automated claims decisions.

Before draftingPackets sorted and deduplicated, wrong-patient material quarantined, page numbers preserved end to end.

While draftingDictate or type against your template. Every statement in the draft carries a page-level citation a reviewer can jump to.

After draftingA qualified reviewer verifies, edits, and signs off. Nothing moves downstream unapproved. See AI medical record summary software, or read how to verify an AI medical record summary.

Browse 41 samples and tools → Chronologies, narrative summaries, expert opinions, IME reports, and more, every one downloadable.

Medical record summary FAQs.

How long should a medical record summary be?There is no fixed standard. A summary should be long enough that every material fact is covered and cited, and short enough that the reader never has to go back into the record to trust it. Length follows record volume and the referral question, not a page quota.

What is the difference between a medical record summary and a medical chronology?A chronology orders every medically significant event by date. A summary condenses the record around the questions the case turns on. Litigation and claims teams often use both: the chronology as the backbone, the summary as the brief on top of it.

Who prepares medical record summaries?Paralegals, nurse reviewers, physicians, and specialized vendors all prepare summaries. What matters in a legal or claims setting is that every material statement traces to a source page and that a qualified reviewer approves the draft before it moves downstream.

Can AI draft a medical record summary?Yes: AI can sort a record, remove duplicates, and draft a summary with page-level citations for human review. Medrecords AI drafts; reviewers verify and decide. No signed opinions, no case-merit scores, and no automated claims decisions come from the software.

What should a medical record summary template include?At minimum: case and patient identifiers, the referral question, an inventory of records reviewed, a treatment chronology, key findings, work or functional status, and flagged gaps or inconsistencies, closed out by a reviewer sign-off.

See a summary of your own file.

Run a representative record through the workflow described above: sorted, deduplicated, drafted with citations, and gated on your reviewer's sign-off.

Book a demo